汉语发音质量评估的实验研究

研究了发音评估系统中通用的置信度测度——后验概率算法,针对它存在的不足,提出了两种改进方案。首先,为了降低计算复杂度,传统算法采用了求最大值算法代替求和算法,在被测发音偏离目标音素集的情况下,这会严重降低后验概率的计算精度,本文提出基于扩展的音素混淆网络的后验概率算法。其次,为使置信度能评估不同语音段长的发音质量优劣,传统算法采用了后验概率的段长规整策略,研究分析发现声学似然值与时间的关系更为紧密,所以本文提出了基于声学似然值的时间规整方案。试验结果表明:与传统算法相比,采用改进的置信度算法能使平均打分错误率相对降低35%左右,有效地改善了计算机辅助语言学习系统的性能。

Experimental investigation of Putonghua pronunciation quality assessment system

As the most effective confidence measure in computer assisted language learning system, the posterior probability is used widely, in which some tricks are applied to reduce the computation complexity. It analyzes the defect of the traditional algorithm and proposes some improvements. First, the traditional algorithm adopts the method of maximum instead of sum in the calculation of the denominator, which seriously reduces the accuracy of posterior probability. Therefore, taking into account both computation complexity and system performance, it proposes an algorithm based on phoneme confusion extended network. Second, in the traditional algorithm, the posterior probability is normalized by its segment time. In fact, the acoustic likelihood is more related with time and grows with the frame number. So, it proposes the acoustic likelihood based normalization algorithm. The experimental results show that compared to traditional algorithm, the proposed algorithm can improve system performance significantly, about 35% average score error rate relatively, and the computation complexity does not increased.